Genomic Footprints of Dryland Stress Adaptation in Egyptian Fat-Tail Sheep and Their Divergence from East African and Western Asia Cohorts Joram M

Animal Science Publications Animal Science

2017 Genomic footprints of dryland stress adaptation in Egyptian fat-tail sheep and their divergence from East African and western Asia cohorts Joram M. Mwacharo International Center for Agricultural Research in the Dry Areas (ICARDA), Addis Ababa, Ethiopia

Eui-Soo Kim Iowa State University

Ahmed R. Elbeltagy Ministry of Agriculture, Cairo, Egypt

Adel M. Aboul-Naga Ministry of Agriculture, Cairo, Egypt

Barbara A. Rischkowsky IFnoterllonawtio thinals Cen andter a fddor Aitgionricualltu wralork Reses aartc:h hintt tps://lhe Dry Aibr.edras.i (asICtAaRteDA.edu/a), Addniss_pubs Ababa, Ethiopia Part of the Agriculture Commons, Molecular Genetics Commons, and the Sheep and Goat See next page for additional authors Science Commons The ompc lete bibliographic information for this item can be found at https://lib.dr.iastate.edu/ ans_pubs/365. For information on how to cite this item, please visit http://lib.dr.iastate.edu/ howtocite.html.

This Article is brought to you for free and open access by the Animal Science at Iowa State University Digital Repository. It has been accepted for inclusion in Animal Science Publications by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Genomic footprints of dryland stress adaptation in Egyptian fat-tail sheep and their divergence from East African and western Asia cohorts

Abstract African indigenous sheep are classified as fat-tail, thin-tail and fat-rump hair sheep. The fat-tail are well adapted to dryland environments, but little is known on their genome profiles. We analyzed patterns of genomic variation by genotyping, with the Ovine SNP50K microarray, 394 individuals from five populations of fat-tail sheep from a desert environment in Egypt. Comparative inferences with other East African and western Asia fat-tail and European sheep, reveal at least two phylogeographically distinct genepools of fat-tail sheep in Africa that differ from the European genepool, suggesting separate evolutionary and breeding history. We identified 24 candidate selection sweep regions, spanning 172 potentially novel and known genes, which are enriched with genes underpinning dryland adaptation physiology. In particular, we found selection sweeps spanning genes and/or pathways associated with metabolism; response to stress, ultraviolet radiation, oxidative stress and DNA damage repair; activation of immune response; regulation of reproduction, organ function and development, body size and morphology, skin and hair pigmentation, and keratinization. Our findings provide insights on the complexity of genome architecture regarding dryland stress adaptation in the fat-tail sheep and showcase the indigenous stocks as appropriate genotypes for adaptation planning to sustain livestock production and human livelihoods, under future climates.

Disciplines Agriculture | Animal Sciences | Genetics and Genomics | Molecular Genetics | Sheep and Goat Science

Comments This article is published as Mwacharo, Joram M., Eui-Soo Kim, Ahmed R. Elbeltagy, Adel M. Aboul-Naga, Barbara A. Rischkowsky, and Max F. Rothschild. "Genomic footprints of dryland stress adaptation in Egyptian fat-tail sheep and their divergence from East African and western Asia cohorts." Scientific epr orts 7 (2017): 17647. doi: 10.1038/s41598-017-17775-3. Posted with permission.

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 License.

Authors Joram M. Mwacharo, Eui-Soo Kim, Ahmed R. Elbeltagy, Adel M. Aboul-Naga, Barbara A. Rischkowsky, and Max F. Rothschild

This article is available at Iowa State University Digital Repository: https://lib.dr.iastate.edu/ans_pubs/365 www.nature.com/scientificreports

OPEN Genomic footprints of dryland stress adaptation in Egyptian fat- tail sheep and their divergence Received: 6 September 2017 Accepted: 20 November 2017 from East African and western Asia Published: xx xx xxxx cohorts Joram M. Mwacharo1, Eui-Soo Kim2, Ahmed R. Elbeltagy3, Adel M. Aboul-Naga3, Barbara A. Rischkowsky1 & Max F. Rothschild2

African indigenous sheep are classifed as fat-tail, thin-tail and fat-rump hair sheep. The fat-tail are well adapted to dryland environments, but little is known on their genome profles. We analyzed patterns of genomic variation by genotyping, with the Ovine SNP50K microarray, 394 individuals from fve populations of fat-tail sheep from a desert environment in Egypt. Comparative inferences with other East African and western Asia fat-tail and European sheep, reveal at least two phylogeographically distinct genepools of fat-tail sheep in Africa that difer from the European genepool, suggesting separate evolutionary and breeding history. We identifed 24 candidate selection sweep regions, spanning 172 potentially novel and known genes, which are enriched with genes underpinning dryland adaptation physiology. In particular, we found selection sweeps spanning genes and/or pathways associated with metabolism; response to stress, ultraviolet radiation, oxidative stress and DNA damage repair; activation of immune response; regulation of reproduction, organ function and development, body size and morphology, skin and hair pigmentation, and keratinization. Our fndings provide insights on the complexity of genome architecture regarding dryland stress adaptation in the fat-tail sheep and showcase the indigenous stocks as appropriate genotypes for adaptation planning to sustain livestock production and human livelihoods, under future climates.

The development of high throughput genome-wide assays and associated computational tools, have made domestic livestock attractive for investigating how an organism’s genome is infuenced by its production and natural environment. Te existence of independent livestock populations/breeds within a species, presents a natural experimental design that can be used to study the genetic mechanisms of adaptive divergence arising from bio-physical and/or anthropological selection. For instance, genome-wide SNP and sequence data has been used to explore genetic mechanisms of adaptation to contrasting environments1–4 and investigate evidence for genomic selection relating to domestication, breed formation and improvement5 in livestock species. Although sheep were domesticated in the Fertile Crescent, Africa is endowed with a diverse repository of the species represented by 179 breeds/populations6 that have been classifed into three broad groups; thin-tail, fat-tail and fat-rump hair sheep. Te thin-tails occur in Sudan and in the sub-humid and humid regions of West Africa. Te fat-tails occur across the deserts of northern Africa, and in the highlands, semi-arid and arid environments of eastern and southern Africa. Te fat-rumps are found exclusively in the semi-arid and arid zones of the Horn of Africa. Te thin-tails are the most ancient in the continent and were introduced, via the Isthmus of Suez and/ or the southern Sinai Peninsula, whereas the fat-tail’s arrived much later, initially via northeastern Africa (Egypt) and later via the Horn of Africa (Ethiopia)7. Te origin of the fat-rumps remains unknown.

1Small Ruminant Genomics Group, International Center for Agricultural Research in the Dry Areas (ICARDA), P. O. Box 5689, Addis Ababa, Ethiopia. 2Department of Animal Science, Iowa State University, 2255 Kildee Hall, Ames, IA, 50011-3150, USA. 3Animal Production Research Institute (APRI), Agriculture Research Centre (ARC), Ministry of Agriculture, Nadi Elsaid Street, Dokki, Cairo, Egypt. Correspondence and requests for materials should be addressed to J.M.M. (email: [email protected])

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 1 www.nature.com/scientificreports/

Figure 1. Population structure of the study populations revealed by (a) DAPC and (b) PCA analyses.

Te genomes of African indigenous sheep have been subjected mainly to natural selection driven by tropical and sub-tropical climates, diseases and parasites. Although their productivity is much lower than that of commercial breeds under intensive production systems, indigenous sheep are ofen the only option available to millions of resource-poor farmers in agro-pastoral and pastoral production systems, where exotic improved genotypes under-perform under limited (quality and quantity) feed and water resources, and high ecto- and endo-parasite and disease challenges. Tis is evident in Egypt, a country within the Sahara desert, where sheep of the fat-tail type, are preferred by livestock keepers because of their excellent adaptation to desert-like conditions8. Tis adaptation and the historical signifcance of Egypt as one of the centers of dispersion of domesticates into Africa, makes Egyptian indigenous sheep of interest in understanding the genetic history of indigenous fat-tail sheep in the continent and the genomic mechanisms underlying their adaptation to dryland environments, which remain poorly investigated. Here, we generated genotype data using the Ovine SNP50K BeadChip from fve populations of Egyptian indigenous sheep representative of the fat-tail hair sheep diversity found across the dry belts of Africa, the Middle East and Asia to investigate their genome profles. Comparative genome analysis with fat-tail sheep from East Africa and western Asia provided insights on the history of the genotype in northeastern and East Africa regions, and the subsequent inclusion of European commercial breeds in selection sweep analysis, allowed us to identify unique genome profles of fat-tail hair sheep to dryland adaptation. Results Population genetic analysis. Population genetic relationships were assessed with PCA and DAPC (Fig. 1) using 5,140 SNPs that were selected to be unlinked. Te frst two principal components of the DAPC and PCA accounted for 7.93% and 9.80% (PC1) and 4.58% and 6.74% (PC2), respectively of the total genetic variation. Te PC1 separated Egyptian and non-Egyptian populations. Te PC2 separated East African and western Asia fat-tail sheep, which seem to cluster together, from the European breeds. Te fve Egyptian populations clustered very close together, but the non-Egyptian ones disperse along the vertical plane of the two plots. Tis suggests a lower level of genetic variation between the Egyptian populations but a comparatively higher one between the non-Egyptian populations. Generally, higher PCs (>2) did not result in observable genetic clusters.

Detection of candidate selection sweep regions. Based on the results of the DAPC and PCA, SNP genotypes were used to estimate allele frequency diferentiation, measured as di, in a pairwise comparison between the Egyptian and non-Egyptian populations. Te genome wide distribution of di values for each SNP is shown in Fig. 2a. A total of 109 signifcant SNPs (di ≥ 4.0) defned seven candidate regions across six chromosomes (Oar1, Oar2, Oar3, Oar8, Oar9, Oar27; Fig. 2a; Supplementary Table S1a). Te strongest candidate region was on Oar9 spanning 24 signifcant SNPs and 15 genes. Te RsB analysis revealed 154 signifcant SNPs (pRsB ≥ 4.0) that defned 10 candidate selection sweep regions across nine chromosomes (Oar1, Oar2, Oar5, Oar7, Oar8, Oar13, Oar15, Oar20, Oar26; Fig. 2b; Supplementary Table S1b). It identifed two candidate regions as the strongest, one on Oar2 and the other on Oar15, spanning 107 and 9 signifcant SNPs, and, 28 and 0 genes, respectively. Te intra-population iHS analysis was performed for the fve Egyptian populations grouped based on the PCA and DAPC. It identifed 47 signifcant SNPs (piHS ≥ 4.0) that defned 14 candidate regions across eight chromosomes (Oar1, Oar2, Oar3, Oar6, Oar10, Oar13, Oar15, Oar17; Fig. 2c; Supplementary Table S1C). Two candidate regions, that were each defned by seven signifcant SNPs, on Oar1 and Oar13 and spanning 46 and 17 genes, respectively were the strongest.

Overlap between candidate selection sweep regions. The three approaches (di, RsB, iHS) used here to detect selection sweeps revealed 31 candidate regions across 15 chromosomes. Selection signatures were identifed on Oar1, Oar2, Oar3, Oar13 and Oar15 by more than one approach. Te signatures identifed by

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 2 www.nature.com/scientificreports/

Figure 2. Manhattan plots for results of selection signature analysis undertaken using the (a) di, (b) RsB and (c) iHS approaches for the study populations.

the three approaches on Oar1 had an overlapping segment (di = 19,517,811–20,118,195 bp; RsB = 19,651,513– 19,761,666 bp; iHS = 19,409,931–21,607,699 bp) (Supplementary Table S1a, S1b, S1c) spanning two genes (TOE1, TESK2). Te signatures identifed by the three approaches on Oar2, and by di and iHS on Oar3, had no overlaps. Similarly, the selection sweeps that were identifed by the three approaches on Oar13 and Oar15 had no overlapping segments. Reducing the signifcance threshold for di to ≥ 3.0, resulted in overlapping segments with iHS on Oar13 (di = 57,868,286–58,232,687 bp; iHS = 57,771,173–58,251,812 bp) that spanned seven genes (PCK1, CTCFL, ENSOARG00000017883, RAE1, SPO11, BMP7, ZBP1), and Oar15 (di = 43,962,190–44,595,229 bp; iHS = 44,112,180–44,332,191 bp) that spanned three genes (ENSOARG00000015306, ENSOARG00000017164, ENSOARG00000017173).

Gene content and functional annotation of the candidate regions. From the 31 candidate selection sweep regions, seven (di = 2, RsB = 3, iHS = 2) spanned no genes (Supplementary Table S1a, S1b, S1c) on the OARv4.0 genome assembly. Such regions have also been reported in cattle5,9,10. We investigated this further by checking the gene content of the seven regions against the Bovine UMD3.1 and Capra hircus V1 (ARS1 (GCF_001704415.1)) genome assemblies. Interestingly they spanned 83 and 18 genes, respectively on the bovine and caprine genomes, suggesting incomplete annotation of the ovine genome assembly. Genome-wide, we identifed 172 genes mapping to 24 (di = 5, Rsb = 7, iHS = 12) candidate regions that were defned by 218 signifcant SNPs across 12 chromosomes (Supplementary Table S1a, S1b, S1c). We performed functional enrichment for the 172 genes using the Enrichr web tool (See Supplementary Table S2). Te rank based and combined score ranking gave similar results and revealed ten GO terms (Table 1) as the most signifcant (P ≤ 0.01). Te genes were associated with diverse biological functions and some had roles in multiple functions. Relevant to this study was that majority of the functions were associated with adaptation to dryland environment stress (Supplementary Table S3a and S3b). Tey included response to feed stress; lipid, protein and carbohydrate metabolism; response to heat/temperature stimulus and oxidative stress; protection from ultraviolet radiation; regulation of immune response, DNA damage repair, transcription and translation, protein modifcation and RNA processing. Other functions included regulation of body size, growth and development; muscle structure, function and adaptation; kidney function and development; and reproduction and nervous system development and function. Comparisons against the KEGG pathway and WikiPathway databases (Supplementary Table S2), revealed two and fve pathways (Table 1), respectively as the most signifcant (P < 0.05). In general, the analysis of GO terms shows an over-representation of GO categories in pathways relating to stress response and which may underlie dryland stress adaptation in the fat-tail sheep (Supplementary Table S4a, S4b, S4c).

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 3 www.nature.com/scientificreports/

Description GO Term P-value Associated genes (a) Enrichment analysis Regulation of stem cell maintenance GO:2000036 P = 0.005938196 TAL1, BMP7 Granulocyte diferentiation GO:0030851 P = 0.005938196 L3MBTL3, TAL1 Vitamin biosynthetic process GO:0009110 P = 0.007613258 AKR1A1, MMACHC Microtubule cytoskeleton organization involved in mitosis GO:1902850 P = 0.005172798 KIF3B, CHEK2 Spindle assembly involved in mitosis GO:0090307 P = 0.003174766 KIF3B, CHEK2 Negative regulation of embryonic development GO:0045992 P = 0.014909626 COL5A2, BMP7 Erythrocyte maturation GO:0043249 P = 0.003790269 L3MBTL3, TAL1 Benzene containing compound metallic process GO:0042537 P = 0.013736768 CMPK1, CYP4B1 Hair cell diferentiation GO:0035315 P = 0.005938196 ERCC3, MYO6 Protein-lipid complex assembly GO:0065005 P = 0.008521348 BIN1, PLAGL2 (b) KEGG Pathway Pyrimidine metabolism hsa00240 P = 0.049048713 POLR2D, CMPK1, DCK Citrate cycle (TCA Cycle) hsa00020 P = 0.022794916 IDH3B, PCK1 (c) WikiPathway ZBP1, ESRP1, ANKAR, GRSF1, KIAA1429, mRNA processing WP310 P = 0.036816137 RAE1, SNRPB TCA Cycle WP434 P = 0.022794916 PDP1, IDH3B WP43/ Oxidation by Cytochrome P450 P = 0.012581695 CYP27C1, CYP4 × 1, CYP4B1 WP1274 Splicing factor NOVA regulated synpatic proteins WP1983 P = 0.042470409 KCNJ6, EPB41L2 miRNA targets in ECM and membrane receptors WP2911 P = 0.040651645 COL3A1, COL5A2

Table 1. Overexpressed GO terms among the 172 candidate genes identifed to be under selection in sheep.

Discussion Occurring predominantly in the Afro-Asiatic drybelts, the fat-tail sheep account for approximately 25% of the global sheep population. Here, we analysed genotype data generated with the Ovine SNP50K Chip to investigate the genome profles and the genetic basis of adaptation to dry environments in fat-tail sheep from a desert environment in Egypt. Te inclusion of fat-tail sheep from East Africa and western Asia allowed us to gain insights on the history of the fat-tails in Africa. Due to their geographic proximity and ancient history of commercial and religious interactions, we hypothesized that the fat-tail sheep from Egypt and western Asia should show a close genetic relationship. However, the DAPC and PCA showed a genetic divergence between the Egyptian and western Asia and East African populations, respectively and a close genetic relationship between the latter. Tis confrms the results of a previous analysis with microsatellites in a large sample of African indigenous sheep that showed a clear divergence between Egypt’s fat-tail Ossimi and its East and southern African counterparts11. Tese results indicate that Egyptian and East African fat-tail sheep represent diferent genetic stocks, and suggest one of two possibilities; an independent introduction of at least two genepools of fat-tail sheep to Africa, or an introduction of one genetic stock that gave rise to two genepools following reproductive isolation and adaptation to diferent eco-climates. We favour the frst suggestion which is in line with archaeological evidence which supports two separate entry points of fat-tail sheep into Africa, initially via northeast Africa (around 7500 and 7000 years ago) and later (around 5000 years ago) via the Horn of Africa12. Te close genetic relationship between the East African and western Asia populations is difcult to explain but we suggest that it may be due to their recent common history. On the other hand, long-term reproductive isolation may explain the divergence of the Egyptian from western Asia populations. Intensive anthropological selection and/or genetic drif arising from low efective population sizes in the European breeds may explain their divergence from the African and western Asia populations. Te genetic divergence revealed by DAPC and PCA informed the grouping of populations into Egyptian and non-Egyptian ones for selection sweep analysis. Te three approaches (di, iHS, Rsb) detected one overlapping candidate region, and two were detected with di and iHS afer relaxing the stringency of di. A modest number of overlapping candidate regions have been reported in several studies4,13–15 and Bahbahani et al.9 reported none. Although coincident signatures that are detected by multiple approaches may provide strong evidence of selec- tion16, their modest occurrence may be due to algorithm diferences17,18. Tis may also explain why results from diferent studies also tend to difer. Terefore, a genomic region that has been identifed by only one approach does not exclude the possibility that it could be under selection19,20. In total therefore, we detected 24 candidate regions, spanning 172 candidate genes, representing signatures of past and/or on-going selection in Egyptian fat-tail sheep. Unsurprisingly, the regions spanned genes that did not concern production, but rather, adaptation traits. Tis is because the target populations have mainly been exposed, over a long time period, to complex interacting biophysical stressors (heat, solar radiation, physical exhaustion, resource scarcity, parasites etc.) which impact ftness. Tis may be the reason why the spanned genes were associated with diverse physiological, molecular and cellular processes and pathways (Supplementary Table S3) emphasizing the importance of multi-functionality for adaptation to dryland environments. Te large number of candidate regions and genes detected is also not surprising; similar fndings have been reported for livestock species from extreme environ- ments4,13,15,21. It reinforces the fact that adaptation is a complex trait that involves many biological processes and quantitative trait loci each having a small and cumulative efect on the overall expression of the phenotype.

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 4 www.nature.com/scientificreports/

Region/Ecology Type Country Breed/Population Sample size Barki 181 Saidi 72 Farafra 62 North Africa (Subtropical dryland) Indigenous Egypt Souhagi 49 AHS 30 Total 394 Ethiopia Menz 34 East Africa (Tropical Highland) Indigenous Kenya Red Maasai 45 Cyprus Cyprus fat tail 30 Afshari 37 Iran Moghani 34 Qezel 35 Western Asia (Subtropical) Indigenous Karakas 18 Norduz 20 Turkey Sakiz 22 Total 196 Bundner Oberlander 21 Swiss White Alpine 21 Switzerland Valais Black Nose 21 Valais Red 21 East Friesian Brown 39 Germany East Friesian White 9 Border Leicester 48 England Dorset Horn 21 North/Central Europe (Temperate) Selected Wiltshire 23 Gallway 49 Ireland Irish Sufolk 55 Scotland Scottish Texel 80 Germany German Texel 43 Old Norwegian Spaelsau 15 Norway Spael coloured 3 Finland Finnsheep 96 Total 565

Table 2. Sample sizes and information on the goat populations and breeds used in the current study.

Energy and nutrient metabolism is vital for herbivores in food scarce environments. Our candidate regions spanned several genes associated with feeding efciency and regulation of metabolism and, GO clusters associated with energy metabolic processes (glycolysis/gluconeogenesis, TCA Cycle, insulin signaling pathway, pyru- vate metabolism etc.). For instance the gene PIK3R3, which has not been reported before, regulates responses to changes in nutritional conditions as well as cellular metabolism and growth22 and through its association with 5′ AMP-activated protein kinase (AMPK), it serves as a metabolic master switch in response to alterations in cellular energy levels23. Indeed, sheep under heat stress decrease dry matter intake and rumination time by up to 76%24 which is related to eating efciency and metabolic processes25. In the drylands most breeds tend to have small body sizes as an adaptation strategy to scarce and poor quality forages and for thermoregulation25. Tis may explain the occurrence, in candidate regions, of genes such as BMP7, MSTN (GDF8) and STIL which regulate adult and embryonic size, growth and development22,26,27. BMP7 also plays a crucial role in renal function and development28,29. Renal vasodilation, transmembrane transport, water-salt metabolism, bicarbonate absorp- tion, water retention and reabsorption are key functions of the renal cortex and central to desert environment adaptation. Prolonged exposure to intense solar and ultraviolet radiation, which are key abiotic stressors in arid environments, can result in ophthalmic and skin conditions. Genes controlling pigmentation of coat and skin30 and eyelids31 and photoreception and visual protection32 have been identifed and reported to ofer protection against solar and UV radiation. None of these reported genes however, occurred in the candidate regions although our study populations had pigmented skins and coats and some such as Barki, Farafra and Souhagi had pigmented eyelids8. Instead, we observed ERCC3 a major nucleotide excision repair (NER) protein. In humans, mutations in ERCC3 result in skin disorders, such as xeroderma pigmentosum, cockayne syndrome and trichothiodystrophy, which result in sensitivity to UV radiation and oxidative stress33–35. Since NER modulates melanocyte stem cell attrition and development of non-pigmented hair36, ERCC3 may be involved in maintaining and regulating the fate and behaviour of melanocyte stem cells and mature melanocytes, and thus the production of melanin which

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 5 www.nature.com/scientificreports/

is responsible for skin, hair and eye colour. Another gene was TGM3 which is widely expressed in skin cells, spe- cifcally keratinocytes and corneocytes. During keratinocyte diferentiation, TGM3 crosslinks structural proteins and lipids in the formation of cornifed cell envelope which provides the barrier function of epidermis against harmful environmental stimuli such as UV radiation37–39. Our candidate regions spanned several genes (PMS1, SPO11, RAD54L, MUTYH, CHEK2, POLR2D, CMPK1) that maintain cellular functions and DNA repair. For instance, MUTYH repairs 8-oxo-G (a mutagenic product of oxidative DNA damage) in the nucleus and mitochondria, via the base excision repair pathway40,41. PMS1 which belongs to the mutL/hexB family of DNA mismatch repair proteins, also forms heterodimers with MLH1, a DNA mismatch repair protein42. Knockdown of CMPK1 was observed to delay DNA repair during recovery from UV damage, suggesting it contributes to the efciency of the DNA repair process43. Tis fnding could be associated with the fact that long term exposure to acute and chronic heat stress enhances the production of free radicals which can result in DNA damage and induce oxidative stress leading to mitochondria damage44–46, apoptosis and necrosis47,48. Terefore the DNA repair system serves to preserve genomic integrity under excessive exposure to UV radiation in dryland environments. Te initiation of stress response involves the activation of the neuroendocrine system to trigger physiological and/or behavioural responses. We found several genes involved in the development of the nervous system and elic- iting response to stress, indicating the importance of the neuroendocrine system in activating stress response in the candidate regions. One of the genes, DMBX1 is involved in brain and sensory organ development49,50 while TP53INP1 is a key cell stress response protein with antioxidant function51,52. Trough its interaction with the ERK1 and p38 mitogen-activated protein kinases, MKNK1 may play a role in responding to environmental stress and control cytokine production and delayed apoptosis53,54. PLAGL2 is also an oxidative stress responding regulator55. Te activation of the neuroendocrine system and initiation of stress response results however, in the chronic production of glucocorticoids and catecholamines, which can dysregulate immune functions56,57, and corticosteroids which possess potent immunosuppressive properties in lymphocytes58. Tis appears to be counterintuitive. However, we observed several genes such as ZBP1, PRDX1, MAST2 and LURAP in the candidate regions that enhance immune functions. Indeed some genes such as MAST2 and PRDX1 have dual roles. By controlling the activities of TRAF6 and NF-kB; MAST2 regulates immune response and acts as the frst responder to harmful cellular stimuli such as stress, free radicals and UV radiation59,60. While PRDX1 may play an antioxidant protective role in cells, it also contributes to antiviral activity of CD8(+) T-cells61,62. Te ZBP1 activates innate immune response by binding foreign DNA, enhances DNA-mediated induction of type I interferons and other genes that activate innate immune responses, as well as, signaling mechanisms underlying DNA-associated antimicrobial immunity and autoimmune disorders63,64. LURAP1 is an activator of the canonical NF-kB pathway and drives the production of proinfammatory cytokines65. Taken together, these fndings suggest that genes evoking cellular stress and immune responses have been the subject of selection in the course of adaptive evolution to dryland environments. Termal stress compromises fertility through a direct efect of hyperthermia on the reproduction axis or through the indirect efect of thermal stress on feed intake to reduce metabolic heat production, leading to changes in energy balance and nutrient availability. Four genes (CTCFL, MAST2, TESK2, SPO11) associated with male reproduction physiology were detected in the candidate regions. CTCFL is a testis-specifc DNA binding protein that forms methylation-sensitive insulators which regulate X-chromosome inactivation, nuclear architecture and transcription66,67. MAST268 and TESK269 play an important role in spermatid maturation during spermi- ogenesis and spermatogenesis, respectively. In mouse, knockout of SPO11 led to meiotic arrest of spermatocytes at zygotene70,71 resulting in sterility in SPO11−/− homozygous male mice and atrophied testes. Since thermal stress compromises fertility through a direct efect of hyperthermia on the reproductive axis or indirectly through the efects of thermal stress on feed intake to reduce metabolic heat production resulting in changes in energy balance and nutrient availability, our result suggests that reproductive success is a key determinant of adaptive ftness in the fat-tail sheep. Te observation that some of the genes spanned by the candidate regions are enriched for the GO term “response to hypoxia (GO:0001666; GO:0071456)” and “HIF-1 (hypoxia inducible factor-1) signaling pathway (hsa04066)” are novel fndings of our study. Te HIF-1 pathway and response to hypoxia plays an important role in cellular response to systemic oxygen levels and has been associated, so far, with adaptation to high altitude environments1–3. We suggest that this fnding could be related to physical exhaustion arising from long-term trekking of long distances in search of feed and water which results in hypoxia-like conditions and oxygen debt in skeletal muscles. In addition two candidate genes (MSTN/GDF8, BIN1) occurred in the candidate regions under selection. MSTN tightly regulates skeletal muscle homeostasis and promotes the survival of muscle syncitia which make up type I (oxidative/slow) or type II (glycolytic/fast) muscle fbers72. Type I fbers are fatigue-resistant and rich in mitochondria and utilize oxidative metabolism to provide a stable and long-lasting supply of ATP73. Tey have been observed in skeletal muscles of endurance athletes and their activation entrains complex pathways that enhance physical73 and racing performance74. Isoforms of BIN1, are important in the formation of transverse tubules which play a role in skeletal and cardiac muscle contractility and relation75,76. Te selection of MSTN, BIN1 and HIF-1 associated genes could have arisen as an adaptive response to endure physical exhaustion arising from long-term long distance walking. In this study, we generated a catalogue of genetic variants in Egyptian fat-tail sheep. Although the studied populations are only a subset of the fat-tail sheep found in Africa, they illustrate that at least two distinct and phylogeographically structured autosomal gene pools defne the genotype in the continent. Genome-wide diferentiation and LD based scans of selection sweeps identifed several candidate genomic regions under selection that spanned several novel and reported genes with key adaptive physiological functions. Tese genes especially those associated with pigmentation, muscle function and the HIF pathway would need further investigation and validation using full genome sequences and expression studies in sheep and model species and further test hypotheses arising from our study.

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 6 www.nature.com/scientificreports/

Material and Methods Animals. Te animals used in this study are owned by farmers. Prior to sampling, the objectives of the study were explained to them in local languages so that they could make an informed decision with regard to providing consent to sample their animals. Blood sampling was performed by a licensed veterinarian following the guide- lines of the General Organization for Veterinary Service (GOVS), Egypt.

Sampling, SNP genotyping and data quality control. Venous blood was collected into EDTA vacutainer tubes from 394 individuals from fve fat-tail sheep populations (Barki, Saidi, Farafra, Souhagi, AHS) in Egypt (Table 2). Te phenotypic characteristics, and socio-economic and cultural signifcance of the populations were described by Galal et al.8. Te animals were sampled at random from farmers’ focks where they are man- aged under transhumant grazing system, and veterinary care and anthropic selection is modest or not practised. Te blood samples were transferred onto Whatman FTATM Classic cards (GE Healthcare UK Ltd) for storage. Genotyping was performed on FTATM spotted blood at GeneSeek Inc (http://www.neogen.com/Genomics/) on the Illumina Ovine SNP50K BeadChip. Similar genotype data from 565 individuals from 15 breeds of European sheep, 196 individuals from seven populations of western Asia fat-tail sheep and 79 individuals from two populations of fat-tail sheep from East Africa (Table 2) were obtained from the Ovine HapMap project (http://www. sheep-hapmap.org). Te European breeds were used to represent genotypes from a contrasting temperate environment. Te East African populations represented fat-tail sheep from a diferent geographic region within Africa and the western Asia ones represented fat-tails from a region close to the centre of domestication. PLINK 1.0777 was used to perform data quality control and processing. A SNP was excluded from the analysis if it failed the following quality criteria; a SNP call rate greater than 95%, SNP genotyping success rate of greater than 90%, a per-SNP minor allele frequency of less than 0.03 and the presence of mapped autosomal loci. Overall, 1,234 individuals genotyped for at least 95% of the SNPs and 51,407 SNPs that passed quality thresholds were available for analysis.

Population analyses. We performed the principal component analysis (PCA) using the adegenet 1.4–2 package78 for R79 to investigate genetic relationships between individuals and populations. To minimize possible confounding efects of linkage disequilibrium on the underlying pattern of population genetic relationship and structure, one in every ten SNPs was sampled and used in PCA. Tis dataset was also used to perform the dis- criminant analysis of principal components (DAPC) with adegenet 1.4–2 package in R, to assess replication of the population clusters and relationships between individuals and populations. Te PCA aims to reduce the total variance in the dataset and identifes the principal components (PCs) that represent population structures based on genetic correlations between individuals. Te DAPC reveals genetic diferences between groups, as best as possible, while minimizing the variation within the groups.

Signatures of selection based on genetic diferentiation. We computed locus-specifc divergence in allele frequencies between groups of populations revealed by DAPC and PCA, using the di statistic80. Te di is a function of unbiased estimates of pairwise values of FST calculated for each SNP between one or a group of population(s) against the other(s). For each of the 51,407 SNPs that passed quality control, the expected value and standard deviation of FST were calculated. Te FST values were then averaged across SNPs contained in non-overlapping sliding windows of 20 SNPs. Te regions under selection were defned if at least two non-overlapping windows passed the distribution threshold, taking the windows with the highest (top 1%) average values of di as the candidate regions80.

Signatures of selection based on linkage disequilibrium (LD). LD based selection sweeps were identifed using the iHS17 and RsB18 approaches. Tese approaches are based on the extended haplotype homozygo- sity (EHH) estimates of LD. Te iHS detects regions with high levels of unexpected EHH within populations relative to neutral expectations and RsB detects the same but between populations. Te analyses were performed using the rehh package81 in R. SNPs under selection were identifed by transforming iHS and RsB values into piHS (piHS = −log10[1 − 2(Φ(iHS) − 0.5)] and pRsB (pRsB = −log10[1 − 2(Φ(RsB) − 0.5)], where (Φ(x)) represents the Gaussian cumulative distribution function. Assuming the values of iHS and RsB are normally distributed under neutrality, the piHS and pRsB can be interpreted as −log10(P-value). For iHS, the intra-population Integrated Haplotype Score (iHS score) was computed for the fve Egyptian populations as a group. For each SNP, we calculated the natural logarithm of the ratio between the integrated EHH for the ancient (iHHA) and derived (iHHD) alleles. We inferred the allelic states of each SNP using two approaches; (i) the SNP with the highest frequency was taken to be the ancestral allele; and (ii) the states were assigned at random following 100 permutations to ensure consistency. Te RsB scores were computed between the Egyptian and non-Egyptian groups of populations. Te integrated EHHS (site-specifc EHH) score for each SNP and group of populations (iES) were calculated, and the RsB statistic between the two groups calculated as the natural logarithm of the ratio between iESEgyptian and iESnon-Egyptian. Haplotypes for iHS and RsB were constructed by phasing the genotyped SNPs using Beagle82. Haplotype frequencies were then calculated using 20 SNP sliding windows that overlapped by fve SNPs. Te SNPs having piHS and pRsB (log10(P-value)) values ≥ 4.0 (P-value < 0.0001), were considered signifcant. Candidate regions were defned if the values for at least two adjacent SNPs were signifcant. In cases where multiple SNPs were signifcant, a distance of 0.5 Kb up- and down-stream of the extreme signifcant SNPs was used to defne the candidate regions.

Functional annotation of candidate regions and genes. The candidate regions were annotated and the associated genes retrieved from the Ensembl Genome Browser 87 using the Sheep Genome Assembly OARv4.0 (GCF_000298735.2). Functional enrichment was performed with Enrichr83 against terms from the

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 7 www.nature.com/scientificreports/

Gene Ontology (GO), KEGG (www.kanehisa.jp/) and WikiPathways (www.WikiPathways.org/index.php/ WikiPathways) databases which annotates groups of genes with dedicated terms. To rank the GO terms that are relevant to the candidate genes, we assessed the signifcance of overlap between the input gene list and the gene sets in each library using the z-score test statistic of the deviation from the expected rank by the Fisher exact test (rank based ranking). Tis test outperforms the Fisher exact test and is comparable to the combined score ranking test83. Te NCBI Pubmed (http://www.ncbi.nlm.nih.gov/pubmed/), GeneCards (http://www.genecards.org/) and UniProt (http://www.uniprot.org/) databases were also used to determine the ®gene functions.

Data availability. Te data generated in this study is available upon request from Eui-Soo Kim (euisoo.kim@ recombinetics.com). References 1. Qiu, Q. et al. Te yak genome and adaptation to life at high altitude. Nature Genet. 44, 946–949 (2012). 2. Dong, K. et al. Genomic scan reveals loci under altitude adaptation in Tibetan and Dahe pigs. PLoS One 9, e110520 (2014). 3. Gorkhali, N. A. et al. Genomic analysis identifed a potential novel molecular mechanism for high-altitude adaptation in sheep at the Himalayas. Sci. Rep. 6, 29963 (2016). 4. Yang, J. et al. Whole-genome sequencing of native sheep provides insights into rapid adaptations to extreme environments. Mol. Biol. Evol. 33, 2576–2592 (2016). 5. Xu, L. et al. Genomic signatures reveal new evidences for selection of important traits in domestic cattle. Mol. Biol. Evol. 32, 711–725 (2015). 6. DAGRIS. Domestic Animal Genetic Resources Information System (ed. Dessie, T., Hanotte, O. & Kemp, S.). International Livestock Research Institute, Addis Ababa, Ethiopia. http://www.dagris.info/ (2017). 7. Muigai, A. W. T. & Hanotte, O. Te origin of African sheep: archaeological and genetic perspectives. Afr. Archaeol. Rev. 30, 39–50 (2013). 8. Galal, S., Rasoul, F.A., Annous, M.R. & Shoat, I. Small Ruminant Breeds of Egypt. In Iñiguez, L., Characterization of Small Ruminant Breeds in West Asia and North Africa Vol 2: North Africa. International Center for Agricultural Research in the Dry Areas (ICARDA), Aleppo, Syria vi+ 196pp (2005). 9. Bahbahani, H. et al. Signatures of positive selection in East African shorthorn Zebu: A genome-wide single nucleotide polymorphism analysis. Sci. Rep. 5, 11729 (2015). 10. Iso-Touru, T. et al. Genetic diversity and genomic signatures of selection among cattle breeds from Siberia, eastern and northern Europe. Anim. Genet. 47, 647–657 (2016). 11. Muigai, A. W. T. Characterization and Conservation of Indigenous Animal Genetic Resources: Genetic Diversity and Relationships of Fat-tailed and Tin-tailed Sheep of Africa. PhD thesis, Department of Biochemistry, Jomo Kenyatta University of Agriculture and Technology, Juja, Kenya (2003). 12. Giford-Gonzalez, D. & Hanotte, O. Domesticating animals in Africa: implications of genetic and archeological fndings. J. World Prehist. 24, 1–23 (2011). 13. Kim, E.-S. et al. Multiple genomic signatures of selection in goats and sheep indigenous to a hot arid environment. Heredity 116, 255–264 (2016). 14. Gouveia, J. J. S. et al. Genome-wide search for signatures of selection in three major Brazilian locally adapted sheep breeds. Livest. Sci. 197, 36–45 (2017). 15. Kim, J. et al. Te genome landscape of indigenous African cattle. Genome Biol. 18, 34 (2017). 16. Ramey, H. R. et al. Detection of selective sweeps in cattle using genome-wide SNP data. BMC Genomics 14, 382 (2013). 17. Voight, B. F., Kudaravalli, S., Wen, X. & Pritchard, J. K. A map of recent positive selection in the human genome. PLoS Biol. 4, 446–458 (2006). 18. Tang, K., Tornton, K. R. & Stoneking, M. A new approach for using genome scans to detect recent positive selection in the human genome. PLoS Biol. 5, e171 (2007). 19. Hohenlohe, P. A., Phillips, P. C. & Cresko, W. A. Using population genomics to detect selection in natural populations: Key concepts and methodological considerations. Int. J. Plant Sci. 171, 1059–1071 (2010). 20. Oleksyk, T. K., Smith, M. W. & O’Brien, S. J. Genome-wide scans for footprints of natural selection. Philos. Trans. R. Soc. London [Biol] 365, 185–205 (2010). 21. Lv, F.-H. et al. Adaptations to climate-mediated selective pressures in sheep. Mol. Biol. Evol. 31, 3324–3343 (2014). 22. Engelman, J. A., Luo, J. & Cantley, L. C. Te evolution of phosphatidylinositol 3-kinases as regulators of growth and metabolism. Nature Rev. Genet. 7, 606–619 (2006). 23. Winder, W. W. & Hardie, D. G. AMP-activated protein kinase, a metabolic master switch: possible roles in Type 2 diabetes. Am. J. Physiol. Endocrinol. Metab. 277, E1–E10 (1999). 24. Monty, D. E., Kelly, L. M. & Rice, W. R. Acclimatization of St. Croix, Karakul and Rambouillet sheep to intense and dry summer heat. Small Ruminant Res. 4, 379–392 (1991). 25. McManus, C. et al. Skin and coat traits in sheep in Brazil and their relation with heat tolerance. Trop. Anim. Health Prod. 43, 121–126 (2011). 26. Wu, M. Y. & Hill, C. S. Tgf-β superfamily signaling in embryonic development and homeostasis. Dev. Cell 16, 329–343 (2009). 27. Chen, D., Zhao, M. & Mundy, G. R. Bone morphogenetic proteins. Growth Factors 22, 233–41 (2004). 28. Gould, S. E., Day, M., Jones, S. S. & Dorai, H. BMP-7 regulates chemokine, cytokine, and hemodynamic gene expression in proximal tubule cells. Kidney Int. 61, 51–60 (2002). 29. Zeisberg, M. et al. Bone morphogenic protein-7 inhibits progression of chronic renal fbrosis associated with two genetic mouse models. Am. J. Physiol. Renal Physiol. 285, F1060–F1067 (2003). 30. Cieslak, M., Reissmann, M., Hofreiter, M. & Ludwig, A. Colours of domestication. Biol. Rev. 86, 885–899 (2011). 31. Pausch, H. et al. Identifcation of QTL for UV-protective eye area pigmentation in cattle by progeny phenotyping and genome-wide association analysis. PLoS One 7, e36346 (2012). 32. Wu, H. et al. Camelid genomes reveal evolution and adaptation to desert environments. Nature Comms. 5, 5188 (2014). 33. Weeda, G. et al. A presumed DNA helicase encoded by ERCC-3 is involved in the human repair disorders xeroderma pigmentosum and Cockayne’s syndrome. Cell 62, 777–791 (1990). 34. Weeda, G. et al. A mutation in the XPB/ERCC3 DNA repair transcription gene, associated with trichothiodystrophy. Am. J. Hum. Genet. 60, 320–329 (1997). 35. Riou, L. et al. The relative expression of mutated XPB genes results in xeroderma pigmentosum/Cockayne’s syndrome or trichothiodystrophy cellular phenotypes. Hum. Mol. Gen. 8, 1125–1133 (1999). 36. Yu, M. et al. Defciency in nucleotide excision repair family gene activity, especially ERCC3, is associated with non-pigmented hair fber growth. PLoS One 7, e34185 (2012). 37. Kalinin, A. E., Kajava, A. V. & Steinert, P. M. Epithelial barrier function: assembly and structural features of the cornifed cell envelope. Bioessays 24, 789–800 (2002).

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 8 www.nature.com/scientificreports/

38. Eckert, R. L., Sturniolo, M. T., Broome, A. M., Ruse, M. & Rorke, E. A. Transglutaminase function in epidermis. J. Invest. Dermatol. 124, 481–492 (2005). 39. Hitomi, K. Transglutaminases in skin epidermis. Eur. J. Dermatol. 15, 313–319 (2005). 40. Pilati, C. et al. Mutational signature analysis identifes MUTYH defciency in colorectal cancers and adrenocortical carcinomas. J. Pathol. 242, 10–15 (2017). 41. Poulsen, M. L. M. & Bisgaard, M. L. MUTYH Associated Polyposis (MAP). Curr. Genomics 9, 420–435 (2008). 42. Smith, C. E. et al. Dominant mutations in S. cerevisiae PMS1 identify the mlh-Pms1 endonuclease active site and an exonuclease 1-independent mismatch repair pathway. PLoS Genet. 9, e1003869 (2013). 43. Tsao, N., Lee, M. H., Zhang, W., Cheng, Y. C. & Chang, Z. F. Te contribution of CMP kinase to the efciency of DNA repair. Cell Cycle 14, 354–363 (2015). 44. England, K., O’Driscoll, C. & Cotter, T. G. Carbonylation of glycolytic proteins is a key response to drug-induced oxidative stress and apoptosis. Cell Death Difer. 11, 252–260 (2004). 45. Lewandowska, A., Gierszewska, M., Marszalek, J. & Liberek, K. Hsp78 chaperone functions in restoration of mitochondrial network following heat stress. Bioenergetics 1763, 141–151 (2006). 46. Song, X. L., Qian, L. J. & Li, F. Z. Injury of heat-stress to rat cardiomyocytes. Chin. J. Appl. Physiol. 16, 227–230 (2000). 47. Rhoads, R. P., Baumgard, L. H., Suagee, J. K. & Sanders, S. R. Nutritional interventions to alleviate the negative consequences of heat stress. Adv. Nutr. 4, 267–276 (2013). 48. Du, J., Di, H. S., Guo, L., Li, Z. H. & Wang, G. L. Hyperthermia causes bovine mammary epithelial cell death by a mitochondrial- induced pathway. J. Termal Biol. 33, 37–47 (2008). 49. Kawahara, A., Chien, C. B. & Dawid, I. B. Te homeobox gene mbx is involved in eye and tectum development. Dev. Biol. 248, 107–117 (2002). 50. Miyamoto, T. et al. Mbx, a novel mouse homeobox gene. Dev. Genes Evol. 212, 104–106 (2002). 51. Shahbazi, J., Lock, R. & Liu, T. Tumor Protein 53-Induced Nuclear Protein 1 Enhances p53 Function and Represses Tumorigenesis. Front. Genet. 4, 80 (2013). 52. Sándor, N. et al. TP53inp1 gene is implicated in early radiation response in human fbroblast cells. Int. J. Mol. Sci. 16, 25450–25465 (2015). 53. Waskiewicz, A. J. et al. Phosphorylation of the cap-binding protein eukaryotic translation initiation factor 4E by protein kinase Mnk1 in vivo. Mol. Cell. Biol. 19, 1871–1880 (1999). 54. Fortin, C. F., Mayer, T. Z., Cloutier, A. & McDonald, P. P. Translational control of human neutrophil responses by MNK1. J. Leukoc. Biol. 94, 693–703 (2013). 55. Guo, Y., Yang, M. C., Weissler, J. C. & Yang, Y. S. PLAGL2 translocation and SP-C promoter activity – a cellular response of lung cells to hypoxia. Biochem. Biophys. Res. Commun. 360, 659–665 (2007). 56. Padgett, D. A. & Glaser, R. How stress infuences the immune response. Trends Immunol. 24, 444–448 (2003). 57. Salak-Johnson, J. L. & McGlone, J. J. Making sense of apparently conficting data: stress and immunity in swine and cattle. J. Anim. Sci. 85, E81–E88 (2007). 58. Cupps, T. R. & Fauci, A. S. Corticosteroid-mediated immunoregulation in man. Immunol. Rev. 65, 133–155 (1982). 59. Chandel, N. S., Trzyna, W. C., McClintock, D. S. & Schumacker, P. T. Role of oxidants in NF-kappa B activation and TNF-alpha gene transcription induced by hypoxia and endotoxin. J. Immunol. 165, 1013–21 (2000). 60. Fitzgerald, D. C. et al. Tumour necrosis factor-alpha (TNF-alpha) increases nuclear factor kappaB (NFkappaB) activity in and interleukin-8 (IL-8) release from bovine mammary epithelial cells. Vet. Immunol. Immunopathol. 116, 59–68 (2007). 61. Sun, C. C. et al. Peroxiredoxin 1 (Prx1) is a dual-function enzyme by possessing Cys-independent catalase-like activity. Biochem. J. 474, 1373–1394 (2017). 62. He, T. et al. Peroxiredoxin 1 knockdown potentiates β-lapachone cytotoxicity through modulation of reactive oxygen species and mitogen-activated protein kinase signals. Carcinogenesis 34, 760–769 (2013). 63. Takaoka, A. et al. DAI (DLM-1/ZBP1) is a cytosolic DNA sensor and an activator of innate immune response. Nature 448, 501–505 (2007). 64. Rathinam, V. A. & Fitzgerald, K. A. Innate immune sensing of DNA viruses. Virology. 411, 153–162 (2011). 65. Jing, Z. et al. Chromosome 1 open reading frame 190 promotes activation of NF-kB canonical pathway and resistance of dendritic cells to tumor-associated inhibition in vitro. J. Immunol. 185, 6719–6727 (2010). 66. Loukinov, D. I. et al. BORIS, a novel male germ-line-specifc protein associated with epigenetic reprogramming events, shares the same 11-zinc-fnger domain with CTCF, the insulator protein involved in reading imprinting marks in the soma. Proc. Nat. Acad. Sci. USA 99, 6806–6811 (2002). 67. Ohlsson, R., Renkawitz, R. & Lobanenkov, V. CTCF is a uniquely versatile transcription regulator linked to epigenetics and disease. Trends Genet 17, 520–527 (2001). 68. Huang, N. et al. A screen for genomic disorders of infertility identifes MAST2 duplications associated with non-obstructive azoospermia in humans. Biol. Reprod. 93: 61, 1–10 (2015). 69. Røsok, O., Pedeutour, F., Ree, A. H. & Aasheim, H. C. Identifcation and characterisation of TESK2, a novel member of the LIMK/ TESK family of protein kinases, predominantly expressed in testis. Genomics 61, 44–54 (1999). 70. Baudat, F., Manova, K., Yuen, J. P., Jasin, M. & Keeney, S. Chromosome synapsis defects and sexually dimorphic meiotic progression in mice lacking Spo11. Mol. Cell 6, 989–998 (2000). 71. Romanienko, P. J. & Camerini-Otero, R. D. The mouse Spo11 gene is required for meiotic chromosome synapsis. Mol. Cell. Proteomics 6, 975–987 (2000). 72. Bradley, L., Yaworsky, P. J. & Walsh, F. S. Myostatin as a therapeutic target for musculoskeletal disease. Cell. Mol. Life Sci. 65, 2119–2124 (2008). 73. Wang, Y.-X. et al. Regulation of muscle fber type and running endurance by PPARδ. PLoS Biol. 2, e294 (2004). 74. Mosher, D. S. et al. A mutation in the myostatin gene increases muscle mass and enhances racing performance in heterozygote dogs. PLoS Genet. 3, e79 (2007). 75. Tjondrokoesoemo, A. et al. Disrupted membrane structure and intracellular Ca2+ signaling in adult skeletal muscle with acute knockdown of Bin1. PLoS One 6, e25740 (2011). 76. Zhou, K. & Hong, T. Cardiac BIN1 (cBIN1) is a regulator of cardiac contractile function and an emerging biomarker of heart muscle health. Sci. China Life Sci. 60, 257–263 (2017). 77. Purcell, S. et al. PLINK: a tool set for whole-genome association and population based linkage analyses. Am. J. Hum. Genet. 81, 559–575 (2007). 78. Jombart, T. & Ahmed, I. adegenet 1.3-1: new tools for the analysis of genome-wide SNP data. Bioinformatics 27, 3070–3071 (2011). 79. R-Development Core Team. R: A Language and Environment for Statistical Computing Version 3.0.1. Vienna, Austria. R Foundation for Statistical Computing (2013). 80. Akey, J. M. et al. Tracking footprints of artifcial selection in the dog genome. Proc. Nat. Acad. Sci. USA 107, 1160–1165 (2010). 81. Gautier, M. & Vitalis, R. rehh: an R package to detect footprints of selection in genome-wide SNP data from haplotype structure. Bioinformatics 28, 1176–1177 (2012). 82. Browning, S. R. & Browning, B. L. Rapid and accurate haplotype phasing and missing data inference for whole genome association studies using localized haplotype clustering. Am. J. Hum. Genet. 81, 1084–1097 (2007). 83. Kuleshov, M. V. et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. 44, W90–W97 (2016).

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 9 www.nature.com/scientificreports/

Acknowledgements We express our gratitude’s to all livestock farmers in Egypt whose animals were sampled and the APRI personnel for their assistance in sampling. Te International Sheep Genomics Consortium (http://www.sheephapmap.org) is acknowledged. Tis project was supported by funding from the College of Agriculture and Life Sciences of Iowa State University, the State of Iowa, Hatch funds, the Ensminger program and the ARC Egypt-ICARDA Collaborative Research Program. ICARDA also thanks the donors supporting the CGIAR Research Program (CRP) on Livestock (formerly Livestock and Fish CRP). This study forms part of our ongoing efforts to understand the adaptation of local indigenous livestock to improve their productivity. Author Contributions J.M.M., A.M.A., B.A.R., M.F.R. designed the study. A.R.E. collected the samples. J.M.M., E.-S.K., A.R.E. analysed the data. J.M.M. wrote the manuscript. All the authors read and approved the manuscript. Additional Information Supplementary information accompanies this paper at https://doi.org/10.1038/s41598-017-17775-3. Competing Interests: Te authors declare that they have no competing interests. Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

SCIentIfIC REPOrtS | 7:17647 | DOI:10.1038/s41598-017-17775-3 10